Identification of product service common and individual demands based on online reviews

Abstract

Online reviews, as the main way of information sharing after purchase, are latent with a large number of customer demands. A systematic method for identifying customers’ common and individual demands of products service based on online reviews is proposed in this paper. Firstly, based on LDA topic model, this method conducts topic mining on the collected review document corpus to filter out keywords and reviews which are related to product service. Then, based on the influence characteristics of reviews, a modified-Louvain algorithm is proposed to divide the word co-occurrence relationship network of product service keywords into communities. Finally, according to the result of community division, customers’ common demands and individual demands of products service can be identified and derived. This demand identification result can provide a more detailed, effective, targeted and predictive information support for product service modularization, so that can facilitate the design, configuration and optimization of the subsequent product service solutions.